By Zack Sanders – Security Expert at APconnections.
Internet security is an increasingly popular and fascinating subject that has pervaded our lives through multiple points of entry in recent years. Because of this infiltration, security expertise is no longer a niche discipline teetering on the fringe of computer science – it’s an integral part. Computer security concerns have ceased to be secondary thoughts and have made their way to the front lines of business decisions, political banter, and legislative reform. Hackers are common subjects in movies, books, and TV shows. It seems like every day we are reading about the latest security breach of a gigantic, international conglomerate. Customers who once were naive to how their data was used and stored are now outwardly concerned about their privacy and identity theft.
This explosion in awareness has, of course, yielded openings for the opportunistic. Companies now know there is a real business need for security, and there are thus hundreds of solutions available to you to improve your security footprint. But most of them are not telling you the truth about how to really secure your infrastructure. They just want to sell you their product – hyping its potential, touting its features, and telling you to install it and – *poof* – you no longer need to worry about security – something those in the industry call “Security Theater.” In many ways, these companies are actually making you less secure because of this sales point. Believing that you can plug in an “all-in-one device” and have it provide you with all of your security controls sounds good, but it’s unrealistic. When you stop being diligent on multiple levels, you start being vulnerable.
Real security is all about two things:
1) Being PROACTIVE.
2) Implementing LAYERED security controls.
Let’s briefly discuss each of these central tenants of best-practice security.
1) Being proactive is key for many reasons. When you are proactive with security, you are anticipating attacks before they start. This allows you to more calmly implement security controls, develop policies, and train staff before a breach occurs. You should be proactive about security for the same reasons you are proactive about your health. Eating well, exercising, and periodically seeing a doctor are all ways to improve your chances of remaining healthy. It doesn’t guarantee you won’t get sick, much in the same way security controls won’t guarantee you won’t get hacked, but it does greatly improve your odds. And if you are not proactive, just like with your personal health, if something does go wrong, it can often be too late to reverse the effects, as most of the damage has already been done.
2) Implementing a layered approach to security is paramount in reducing the odds of a successful attack. The goal is to take security controls that complement each other on different levels of your infrastructure and piece them together to form a solid line of defense. If one control is breached, another is there to back it up in a different, but equally effective way. It is actually possible to take products that are relatively ineffective on their own (say 75% effective), and layer them to lower the chances of a successful attack to less than 1%. If you implement just four 75%-effective tools, say, check out what your breach success rate becomes: (.25 * .25 * .25 *.25) = .0039 * 100 = 0.39%! That’s pretty impressive!
Here is an analogy
Think of your sensitive data as crown jewels that are stored in the center of a castle. If your only security control is a moat, it wouldn’t take much ingenuity for a thief to cross over the moat and subsequently steal your jewels. One thing we can do to improve security is better our moat. Let’s add some crocodiles – that will certainly help in thwarting would-be crossers. But, even though we’ve beefed up the security of the moat, it’s still passable. The problem is that we can never 100% secure the moat from thieves no matter what we do. We need to add in some complementary controls to back up the security of the moat in case the moat fails. So, we’ll place archers at the four corner towers and install a big door with multiple locks and guards at the front gate. We’ll move the jewels to the cellar and place them under lock and key with a designated guard. Knights will be trained to spot thieves, and there will be a checkpoint outside the castle for all incoming and outgoing guests. Now, instead ofhaving to just cross the moat, a thief would also have to get through the heavy door, through the locks, past the guards, past the archers, into the cellar, past another guard, and into the locked room. On exit, he’d have to get through all these again, including a manual search at the checkpoint. That seems tough to do compared to just crossing the moat.
Your web security infrastructure should work the same way. Multiple policies, devices, and configurations should all work in harmony to protect your sensitive data. When companies are trying to sell you an all-in-one security device, they are essentially trying to sell you a very robust moat. It’s not that their product won’t provide value, but it needs to be implemented as part of an overall security strategy, and it should not be solely relied upon.
How Real Attacks Occur
We have thought a lot lately about exactly how real attacks occur in the wild for organizations with interactive web applications. This is slightly simplistic, but it really seems to boil down to two key origins:
1) A hack results from an AUTOMATED scan or probe.
This is by far the most common type of attack, despite it not being as popular as the other. Many organizations don’t take this type of attack as seriously as they should. They think that just because they are a small, non-influential site with little customer data that they won’t be targeted. And they are probably right – a human attacker won’t be targeting them. But a robot has no discretion. The robot’s goal is to increase hosts in their botnet (for DOS attacks, sending SPAM, etc.), and to siphon off any available sensitive data from the server. The botnets are constantly scouring the Internet, rapidly attempting breaches with known, common patterns. They don’t get too sophisticated.
2) A hack results from a TARGETED attack.
The media has hyped this into the most popular type of attack, but it is much less common. Targeted attacks can begin from multiple motivations. Sometimes, a targeted attack will occur due to interesting results from an automated scan (as in #1, above). The other type of targeted attack is the most dangerous – an attacker, or group of attackers, specifically targeting your site for financial or political reasons. Despite what other products might profess, there is no one-stop solution for stopping this type of attack. A layered approach to security, as discussed above, is key.
Approaches to Dealing with Botnets/Malnets and other Automated Attacks
Botnets are large, distributed networks of private computers and servers that are infected with malicious software without the owner of the system being aware. The botnet computers can be used to scan targets for vulnerabilities or send out SPAM/malicious emails. Using systems registered to someone else provides a layer of anonymity to the attacker. He/she also has increased processing power and resources available at their disposal. Botnets rely heavily on attempting simple intrusions and speed. They often are brute forcing directory listings or credentials and once they’ve exhausted their lists, they move on.
There are a few things you can do to greatly lower the effectiveness of a botnet:
1) Think about if your website really needs to be open to the entire Internet. Are there countries/subnets that you will never receive business from? Why not just block these IP ranges right off the bat? It seems harsh at first, but if you think about it, there is a lot of added security value here for the small risk you turn away a legitimate customer.
2) Implement a tool that monitors the amount of requests received over a given time frame. A normal user won’t ever be requesting pages at the same rate as a botnet. If the request count reaches past a certain threshold, you can confidently block the offending IP.
3) Implement a tool that monitors logs for multiple 404 (Page Not Found) requests. Brute-force tools will generate plenty of 404 requests when they are hammering your servers. If you see multiple 404’s over a short period of time from the same IP, chances are good they are acting maliciously.
4) Look for common patterns in logs that suggest malicious intent. The information discovery process is very important for an attacker (or botnet). It is during this phase that they learn about possible vulnerabilities your sites might have. In order to find these holes, the attacker has to experiment with the site to see how it responds to malicious code. If you can isolate these probing attempts right off the bat, you stand a good chance at cutting off the information gathering process before they get results on potential attack vectors.
5) Implement a file integrity monitoring tool on your web server and have it actively alert to changes in files that are not supposed to change often. If an attacker finds an entry point, one of the first things they will try and do is upload a file to the server. Getting a file to the server is a huge accomplishment for an attacker. They can upload PHP or ASP files that act as shell interfaces to the server itself, and from there can wreak whatever havoc they’d like. With a file integrity monitoring tool, you can know if an file is added within minutes of upload and can deal with the threat before it is wide spread.
NetGladiator is a next-generation Intrusion Prevention System (IPS) made by APconnections that deals with some of the issues above and was built based on how attacks actually occur. It can be an effective layer in your security profile to help block unwanted web-based requests (either from a botnet or a targeting attacker) – you can think of it as a firewall for your web applications. In addition to handling web requests, it can detect time-based anomalies and block IP ranges by country and/or subnet.
NetGladiator has two primary goals:
1) Make your web infrastructure INVISIBLE and UNINTERESTING to probing botnets.
2) Provide value as a LAYERED appliance in case of a targeted attack.
NetGladiator also has some of the following aspects that set it apart from more expensive, overly robust IPS’s:
Unlike other IPSs with insanely robust pattern sets, NetGladiator lets you pick and choose the patterns you’d like it to hit on. Other products inspect for every vulnerability known to man. While this sounds good, it isn’t very practical and often leads to broken functionality, false positives, and total reliance.
Support From a White Knight (a.k.a Professional Hacker)
As part of your support agreement when you purchase a NetGladiator, a real, white knight will help you set up and configure your machine to meet your needs. This includes identifying and patching any existing holes prior to your installation, deciding what issues you might face from a real attacker, and writing you a custom configuration for your box. That’s something that no one else provides – especially at this price point. And, if you want further security assessments performed, additional support hours can be purchased.
Plug and Play
If you’ve set up a NetEqualizer in the past, you’ll find NetGladiator’s installation process to be even easier. Just put it in front of your web servers, cable the box correctly, and turn it on. Traffic will be passing through it instantly. Now all that’s left is to configure your patterns. NetGladiator comes with default patterns in case no customization is necessary. NetGladiator also runs on its own system, and does not require any installs to your web server. This makes it platform independent and will create zero conflicts with your existing software and hardware.
But remember, protecting web applications is just one piece of the puzzle. In order to layer NetGladiator into your overall security strategy, you should complement its use with other controls. Some examples would be:
– Well-defined user and staff policies that deal with insider threats and social engineering
– Full or column-level database encryption
– File integrity monitoring
– Hardware firewalls
– A security assessment by an expert
Need help instituting a layered security strategy? We have experience in all these levels of security controls and are happy to help with NetGladiator implementation or other security-related tasks. Just let us know how we can be of service!
Have some questions about NetGladiator or web security in general? Visit our website, leave a comment, or shoot us an email at email@example.com. In the next blog post, we’ll answer those questions and also discuss common ones we’ve received from customers so far.